Characterization of large and small-plaque variants in the Zika virus clinical isolate ZIKV/Hu/S36/Chiba/2016

An Asian/American lineage Zika virus (ZIKV) strain ZIKV/Hu/S36/Chiba/2016 formed 2 types in plaque size, large and small. Genomic analysis of the plaque-forming clones obtained from the isolate indicated that the clones forming small plaques commonly had an adenine nucleotide at position 796 (230Gln in the amino acid sequence), while clones forming large plaques had a guanine nucleotide (230Arg) at the same position, suggesting that this position was associated with the difference in plaque size. Growth kinetics of a large-plaque clone was faster than that of a small-plaque clone in Vero cells. Recombinant ZIKV G796A/rZIKV-MR766, which carries a missense G796A mutation, was produced using an infectious molecular clone of the ZIKV MR766 strain rZIKV-MR766/pMW119-CMVP. The plaque size of the G796A mutant was significantly smaller than that of the parental strain. The G796A mutation clearly reduced the growth rate of the parental virus in Vero cells. Furthermore, the G796A mutation also decreased the virulence of the MR766 strain in IFNAR1 knockout mice. These results indicate that the amino acid variation at position 230 in the viral polyprotein, which is located in the M protein sequence, is a molecular determinant for plaque morphology, growth property, and virulence in mice of ZIKV.

infection also caused Guillain-Barré Syndrome (GBS) in some patients during the outbreak in French Polynesia in 2013-2014 16,17 . The World Health Organization (WHO) declared ZIKV epidemics in the Pacific countries and territories and the Americas, and its associated complications a public health emergency of international concern (PHEIC) on February 1st, 2016 though the PHEIC was lifted in November 2016.
ZIKV genome is composed of a single-stranded, positive-sense RNA, which encodes 3 structural proteins (C, prM, and E) and 7 nonstructural proteins (NS1, NS2A, NS2B, NS3, NS4A, NS4B, and NS5) in one open reading frame. ZIKV is genetically divided into 2 lineages, African and Asian/American lineages. The recent epidemics in Pacific regions and the Americas were caused by the introduction of Asian/American lineage ZIKV 18,19 .
Twelve imported cases of ZIKV infections have been reported in 2016 in Japan. Recently, we isolated a ZIKV strain (ZIKV/Hu/S36/Chiba/2016) from serum of a patient, who returned from Fiji to Japan in April 2016 38 . Phylogenetic analysis indicated that the isolate belonged to Asian/American lineage ZIKV and showed the highest homology to the Tonga strain isolated in 2016 38 . During the characterization process of the ZIKV/Hu/S36/ Chiba/2016 isolate in vitro, it was revealed that the isolate included at least two variants with different phenotypes. The characteristics of these two variants in terms of plaque formation, growth capacity in vitro, and virulence were elucidated. In this study, we established a reverse genetics system for ZIKV using the MR766 strain 39 . The genetic determinant that caused the difference in phenotypes was determined using the reverse genetics system.

Results
ZIKV/Hu/S36/Chiba/2016 isolate contains two types of ZIKV. The virus was isolated by adding diluted serum of the patient to the culture medium of Vero cells and subsequent blind passage of the culture supernatant to Vero cells. The ZIKV/Hu/S36/Chiba/2016 isolate at this stage was defined as "ZIKV/Hu/S36/Chiba/2016-Vero2", and it formed morphologically different types of plaques, which were large and small in size, in Vero cells (Fig. 1a). The ratio of the number of the large to small plaques was approximately 1:1. When a small aliquot of the ZIKV/Hu/S36/ Chiba/2016-Vero2 isolate was amplified in Vero cells, the amplified virus was defined as "ZIKV/Hu/S36/Chiba/2016-Vero3". Most of the plaques formed by ZIKV/Hu/S36/Chiba/2016-Vero3 were large in size (Fig. 1b). The complete nucleotide sequence of the ZIKV/Hu/S36/Chiba/2016-Vero2 and -Vero3 isolates revealed that there was only one nucleotide difference between these viruses at position 796, which corresponds to position 230 in the viral polyprotein and is located in the structural M protein coding region. The major and minor nucleotides at position 796 in ZIKV/Hu/ S36/Chiba/2016-Vero2 were adenine and guanine, respectively (Fig. 1c). By contrast, the major and minor nucleotides at the same position were guanine and adenine in ZIKV/Hu/S36/Chiba/2016-Vero3, respectively (Fig. 1d). Moreover, the adenine-to-guanine nucleotide change at 796 (A796G) led to an amino acid change at position 230, Gln to Arg (Gln230Arg). To clarify whether the sequence difference caused the alteration in plaque morphology, ZIKV clones were obtained from the ZIKV/Hu/S36/Chiba/2016-Vero2 solution using a limiting dilution method. Nineteen clones were obtained. The complete nucleotide sequences of 10 out of 19 clones were determined (Table 1). Several nucleotide variations were seen among the clones. All 5 clones forming small plaques had an adenine nucleotide at position 796, but all 4 clones forming large plaques had a guanine at the same position. One clone (#28), which was a mixed population of large-plaque and small-plaque viruses, had a mix of adenine and guanine at that position. These results suggest that the amino acid at position 230 was associated with the difference in plaque formation.
In vitro growth of the large-plaque and small-plaque variants of ZIKV. To examine the growth properties of large and small-plaque ZIKV, two clones, designated as ChibaS36#3LP (large plaque) and ChibaS36#8SP (small plaque), were selected from the 19 clones (Fig. 2a). On the one hand, ChibaS36#3LP grew faster than ChibaS36#8SP in Vero cells (Fig. 2b). Infectious titer in the culture supernatant reached 5 × 10 7 PFU/ mL at 72 hours after inoculation, and ChibaS36#3LP titer was over 20-times higher than that of ChibaS36#8SP. On the other hand, both clones showed slower replication kinetics in mosquito C6/36 cells than in Vero cells, and there was less difference in growth rate between the clones (Fig. 2c).

Generation of a recombinant ZIKV G796A/rZIKV-MR766.
To demonstrate whether the amino acid difference at position 230 was associated with plaque formation and growth capacities in Vero cells, a recombinant ZIKV with or without the mutation at position 796 (G796A) was generated by using a reverse genetic system. A reverse genetics system for ZIKV was established by constructing an infectious molecular clone of the ZIKV MR766 strain rZIKV-MR766/pMW119-CMVP (Fig. 3a). Vero cells were transfected with the plasmid, and the culture supernatant fluid was recovered. The recombinant virus nucleotide sequence had no mutation, and the plaque size and the growth kinetics of the recombinant virus were identical to the parental virus, indicating that the reverse genetics system for producing infectious recombinant ZIKV was established (Figs 3b and S1).
ChibaS36#3LP and MR766 strain possessed a guanine at the position 796; therefore, a guanine-to-adenine substitution (G796A) was introduced at position 796 on the genomic region of the MR766 clone to construct a mutant clone G796A/rZIKV-MR766/pMW119-CMVP. The mutant recombinant virus G796A/rZIKV-MR766 was produced successfully by transfecting Vero cells with the mutant clone.
Growth kinetics and plaque formation of recombinant ZIKV G796A/rZIKV-MR766 and parental MR766 virus. The plaque size of G796A/rZIKV-MR766 was clearly smaller than that of the parental strain ( Fig. 4a). Furthermore, the growth kinetics of G796A/rZIKV-MR766 was also significantly slower than that of the parental strain in Vero cells, indicating that the G796A mutation alone caused reductions in plaque size and growth properties in Vero cells (Fig. 4b). However, the parental and G796A mutant viruses grew similarly in C6/36 cells, although there was a slight difference between the two viruses at day 4 ( Fig. 4c).

Pathogenesis of G796A/rZIKV-MR766 and parental MR766 virus in IFNAR1-KO mice. Previous
reports showed that interferon alpha/beta receptor 1 gene-knockout (IFNAR1-KO) mice were vulnerable to MR766 infection 30 . Using the mouse model, we evaluated the virulence of recombinant G796A/rZIKV-MR766 (Fig. 5). All the mice infected with 1 × 10 2 PFU of the MR766 strain died within 8 days. However, the survival of mice infected with G796A/rZIKV-MR766 from inoculation was longer than that of mice infected with MR766, and 5 out of 6 mice inoculated with recombinant G796A/rZIKV-MR766 died within 11 days and one survived. Statistical analysis (log-rank test) indicated that there was a significant difference in the survival rates between the two groups (P = 0.033). Body weight loss of mice inoculated with G796A/rZIKV-MR766 was also smaller than that of mice inoculated with the parental MR766 virus ( Figure S2). The viremia level in the mice inoculated with G796A/rZIKV-MR766 was lower than that in the MR766-inoculated mice ( Figure S3).

Nucleotide sequencing at position 796 of ZIKV/Hu/Chiba/S36/2016 from the patient serum and the virus isolated using C6/36 cells.
To confirm whether the nucleotide at position 796 in ZIKV sequence was mainly adenine or guanine, the nucleotide sequence was directly determined from the serum sample (Table 2). Guanine was detected at that position in the serum sample, suggesting that the majority of ZIKV in the patient serum are large-plaque variants. We also amplified ZIKV/Hu/S36/Chiba/2016-Vero2, which was then propagated in C6/36 cells three times (ZIKV/Hu/S36/Chiba/2016-Vero2-C6/3); the nucleotide sequence of the virus was then determined ( Table 2). Both adenine and guanine at position 796 were detected, suggesting that the adenine-type and guanine-type viruses coexisted in the culture supernatant fluid of C6/36 cells infected with ZIKV/Hu/S36/Chiba/2016-Vero2-C6/3.

Discussion
In this study, we demonstrated that the viral solution of the ZIKV/Hu/S36/Chiba/2016-Vero2 isolate contained at least two phenotypes of ZIKV with different plaque size formation and growth kinetics in vitro and that a nucleotide at position 796 (amino acid residue at position 230 in the viral polyprotein), which locates in M protein, was responsible for the differences. We also proved that the G796A mutation decreased the virulence of ZIKV MR766 strain in IFNAR1-KO mice.
Small-plaque variants were found in low-passaged viral solutions and gradually disappeared from the culture supernatant by repeated passages in Vero cells (Fig. 1). Therefore, it was speculated that these small-plaque variants represented the largest population in the patient serum. However, the sequencing data indicated that most viruses in the patient serum consisted of the large-plaque variant, because the nucleotide detected was mainly guanine at position 796. How the small-plaque variant was isolated by inoculating Vero cells with the serum samples, as well as the large-plaque variant, remains to be elucidated. Both the large and small-plaque variants were found after passaging the isolate (Vero2) in C6/36 cells ( Table 2). The growth capacity of ZIKV/Hu/S36/ Chiba/2016 in C6/36 cells was clearly lower than that in Vero cells, but no clear difference in growth rate was observed between the large and small-plaque clones (Fig. 2). These results suggest that the small-plaque variant may disappear more slowly in C6/36 cells than in Vero cells through multiple passages. The difference in growth kinetics between the parental MR766 strain and the G796A mutant virus was also small in C6/36 cells, indicating that the amino acid at position 230 is not important for the ZIKV growth in C6/36 cells.
In flaviviruses, prM protein is a membrane glycoprotein and forms a heterodimer with envelope E protein on the surface of virus particles. The roles of prM protein are to assist in the proper folding of E protein and to prevent E protein from undergoing acid-catalyzed rearrangement to the fusogenic form during the secretory pathway. In the maturation process of virion, prM is cleaved into pr and M regions by a cellular protease furin. M region of ZIKV consists of 75 amino acid residues (about 10 kDa) and an N-terminal loop, a membrane-associated helix, and two transmembrane (TM) domains (Fig. 6) 40 . The amino acid at position 230 is located on the loop region. The TM and helix domains are deemed to be involved in assembly and maturation of virus particles and dimerization with E protein [41][42][43] . However, there is little information about the N-terminal loop. In this study, we showed that the amino acid substitution in the loop influences ZIKV growth in vitro and its virulence in vivo. This raises the possibility that the region plays an important role in ZIKV replication in vitro and in vivo in mammalian cells. The N-terminal loop is considered to lie within the hole in between the homodimers of E ectodomain 44,45 , suggesting that the loop may also be associated with the formation of ZIKV particles. Recent report demonstrated that a single mutation in the ZIKV prM causes fetal microcephaly in mice, suggesting that ZIKV prM is involved in the pathogenicity of ZIKV 46 . The African lineage MR766 strain was passaged in mouse brain repeatedly and has adapted to mice, in which it was detected in various tissues, such as blood, spleen, liver, kidney, brain, spinal cord, placenta, testis, and body fluids, and led to dysfunction of placenta and testis, death and reduced proliferation of neural progenitor cells, and fetal brain infection in the animal models 47 . Our results of in vitro analyses suggest that the G796A mutation influenced ZIKV growth in IFNAR1-KO mice and in vivo, the growth rate of the mutant G796A/rZIKV-MR766 virus in the mice was lower than that of the parental MR766 virus. We established the reverse genetics of ZIKV using the MR766 strain in the present study. Using this system, we could demonstrate clearly that adenine at position 796 in the ZIKV genome was involved in small-plaque formation and lowered growth capacity in vitro and in vivo. Thus, the reverse genetics system that we developed enables us to understand of the nature of ZIKV in more details.

Methods
Viruses, cell culture, plaque assay, and virus cloning. The ZIKV strain ZIKV/Hu/S36/Chiba/2016 strain 38 and the MR766 strain (MR766-NIID, accession no. LC002520), originally isolated from a rhesus monkey in the Zika forest of Uganda in 1947 and maintained in the National Institute of Infectious Diseases (NIID), Japan, were used in this study. These viruses were propagated in Vero cells (strain 9013) and mosquito-derived C6/36 cells. Vero cells and C6/36 cells were cultured at 37 °C and 28 °C, respectively, in 5% CO 2 in Eagle's minimum essential medium (MEM) supplemented with 10% heat-inactivated fetal bovine serum (FBS) and 100 U/mL penicillin-streptomycin. To determine the infectious titer of the virus, a plaque assay was performed as described previously 48 . Briefly, Vero cells (2 × 10 5 cells/well) were plated in 12-well plates and inoculated with serially diluted viruses. Four (Fig. 3) or five (Figs 1, 2 and 4) days after inoculation, the cells were fixed with a 3.7% (v/v) formaldehyde solution in phosphate-buffered saline and then were stained with a methylene blue solution for 2 hours. For cloning viruses, the ZIKV/Hu/S36/Chiba/2016-Vero2 solution, which was obtained by inoculating the patient serum into Vero cells and subsequent passage of the supernatant in Vero cells 38 , was diluted with MEM supplemented with 2% FBS (MEM-2% FBS) and then inoculated into Vero cells in 24-well plate at 0.5-1.0 infectious virus particle/well. Seven days after inoculation, the supernatants were recovered and then used to check the plaque size of the clones using a plaque assay as described above. Ten out of the 19 clones obtained were used for determination of the complete nucleotide sequence. Genome sequencing. Viral RNA was isolated from culture supernatants and serum of the Zika patient 38 using High Pure Viral RNA Kit (Roche Diagnostics, IN, USA) and was used for synthesis and amplification of cDNA using a one-step RT-PCR method using PrimeScript II High Fidelity One Step RT-PCR kit (TAKARA Corp., Shiga, Japan). Viral cDNA were amplified by PCR using the primer sets described in Supplementary  Table 1. PCR products were sequenced using the dideoxy sequencing reaction kit (BigDye terminator method; Thermo Fisher Scientific, Waltham, MA, USA) and Genetic Analyzer 3500 (Thermo Fisher Scientific) with primers specific for ZIKV. The raw sequencing data were assembled to reconstruct the complete ZIKV genome and translated into amino acid sequences using GENETYX gene analysis software (Genetyx Corp., Tokyo, Japan).

Production of recombinant ZIKV.
A plasmid pMW119-CMVP, which contains a CMV promoter at 5′, and ribozyme and polyA tail sequence at 3′ of the viral sequence 49 , was amplified with the primers pCMV. del.f and pCMV.del.r (Supplementary Table 1) and digested with the restriction enzyme Dpn I. MR766-NIID RNA was extracted and used for the synthesis of viral cDNA using ReverTra Ace (TOYOBO, Osaka, Japan). Eight regions (A to H regions), which included the overlapped sequences adjacent to the other regions, were amplified by PCR using Q5 Hot Start High-Fidelity Master Mix (New England Biolabs, Ipswich, MA, USA) and the primers (Supplementary Table 1 and Fig. 3a). The fragments were mixed and then amplified by PCR using 5′-terminal and 3′-terminal primers, pCMVIF.ZKV001f, and pCMVIF.ZKV10794r (Supplementary Table 1), to synthesize the complete MR766-NIID cDNA. The full-length cDNA and the amplified pMW119-CMVP were ligated using In-Fusion HD (TaKaRa-Clontech, Shiga, Japan). Complete plasmids were amplified in NEB Stable Competent cells (New England Biolabs) and then recovered. Vero cells were transfected with the recombinant clone rZIKV-MR766/pMW119-CMVP using X-tremeGene HP (Roche) transfection reagent. After incubation of the transfected cells for 4 days, the culture supernatant was collected and inoculated onto Vero cells. Four days after inoculation, the culture supernatant was recovered, and its virus titer was measured with a plaque assay. A  recombinant molecular clone, G796A/rZIKV-MR766/pMW119-CMVP, was constructed using the ZIKV strain MR766-NIID backbone. A guanine nucleotide at position 796 of rZIKV-MR766/pMW119-CMVP was replaced with an adenine nucleotide by site-directed mutagenesis (inverse PCR method) using primers RtoQ.f and RtoQ.r (Supplementary Table 1). The recombinant virus was recovered by transfection of Vero cells with the recombinant clone plasmid described above. Complete nucleotide sequence of the recombinant virus was determined, and it was confirmed that no spontaneous mutations were present on the genomes.
Growth kinetics analysis. Cells were plated in a 6-well culture plate (3 × 10 5 cells/well for Vero and 6 × 10 5 cells/well for C6/36 cells) and infected with ZIKV, including recombinant ZIKV, at a multiplicity of infection of 0.01 plaque-forming units (PFU) per cell. Small aliquots of the media were recovered periodically, and the titer of the aliquots was determined by a plaque assay on Vero cells grown in 12-well culture plates. To evaluate the plaque size, Vero cells (3 × 10 5 cells/well) were plated in 12-well plates and inoculated with the viruses. Five days after inoculation, cells were fixed and then stained with a methylene blue solution as described above. Statistical comparison of growth curves was performed using Prism (GraphPad Software, La Jolla, CA, USA) that uses the two-way ANOVA test.

Mouse challenge.
Mouse experiments were performed in accordance with the "Guidelines for animal experiments performed at National Institute of Infectious Diseases" under the approval (no. 116067) from the Animal Welfare and Animal Care Committee of the National Institute of Infectious Diseases, Japan. IFNAR1-KO C57BL/6 mice were provided by the RIKEN Bio Resource Center, Japan, through the National Bio-Resource Project of the MEXT, Japan 50,51 . All mice were bred and maintained in a specific-pathogen-free environment. Mice (7-9 weeks old, n = 6) were inoculated with 1 × 10 2 PFU of each virus solution diluted with MEM-2% FBS through the subcutaneous route in the footpad. The mice were observed clinically, and their body weight was measured every day for 2 weeks after inoculation. Survival curve comparisons were performed using JMP13 software (SAS Institute Japan, Tokyo, Japan) statistical analysis with a log-rank test. Blood was collected after inoculation of the viruses at an interval of 2 days (3 mice, day 1, 3, and 5; the rest 3 mice, day 2, 4, and 6), and TCID50 was calculated as described previously 52 . The dataset are available from the corresponding author on reasonable request.